Towards automatic building extraction: Variational level set model using prior shape knowledge

摘要

A novel variational level set model for multiple-building extraction from a single remote image is proposed in this paper. Multi-competing shapes are considered together with the level set model, the curve evolution is constrained by the prior shape knowledge and the label function that dynamically indicates the region with which the prior shape should be compared. The building extraction is addressed through a level set image segmentation approach that involves the use of the label function, as well as the prior shape knowledge. In addition, the proposed model permits translation, scaling, and rotation of the prior shape. Experimental results and the qualitative and quantitative evaluations demonstrate the potential of the approach.